Using GUDMAP - Tutorial

This tutorial illustrates how the GUDMAP website can be used to search gene expression data to find candidate genes that may play a role in a known disease.
(View this tutorial as pdf)

The starting point is a known disease – Renal Adsyplasia

We want to know a few things:
1. What genes are associated with this disease?
2. Where are genes associated with this disease expressed in the kidney?
3. What other genes share similar expression patterns to these genes?

1. Using GUDMAP Disease Resource
Navigate to Disease > Query Disease-Gene Associations

Use the search box to select or type in our disease of interest. The list of disease names is taken from OMIM. Using this query we can find if there are known genes associated with our disease of interest.

Figure 1. Query Disease-Gene Associations

2. Select Gene from Results Page

The results page will return a list of genes that have an association to the disease. Associations flagged as being NCBI = 1 are taken from NCBI’s list of disease-gene associations (mim2gene -  In situ data = 1 means there is in situ data for that gene in GUDMAP.

We want to look at a gene that has in situ data so that we can see where it is expressed in the kidney. In this result set both RET and UPK3A have mutations that cause Renal Adsyplasia and there is in situ data in GUDMAP for these genes. So we select one of these (Upk3a) to investigate further. Click on the Upk3a link in the Mouse Gene Symbol column.

Figure 2. Disease-Gene Results Page

3. Upk3a – Gene Page – access to expression data

All information about Upk3a can be viewed on its gene page, including details of what entries have been submitted to the database for Upk3a

Figure 3. Upk3a Gene Page

The In Situ section of the page shows how many in situ hybridisation (ISH) entries GUDMAP holds for this gene – in this case 5 wholemount ISH and 5 section ISH entries. Each in situ can be viewed by clicking on the GUDMAP ID in the In-Situs section (see Fig. 3).

We want to know where Upk3a is expressed in the kidney and use this information to search for other genes that share a similar expression pattern. So we want to look at a section ISH entry for the metanephros – in this case GUDMAP:9904. Click on the ID number to view the entry page.

4. View Expression Mapping

Figure 4. Expression Mapping of GUDMAP:9904. Annotations viewed as a list.

On the entry page for GUDMAPP:9904 there is a section called ‘Expression Mapping’. This gives details of all the annotation for that entry. By default they are displayed as an ontology tree. The annotation can be viewed as a simple list (Figure 4) by clicking on the appropriate tab. Now we have a discrete list of annotations for GUDMAP:9904.
We can use these annotations to perform a search of the database to find genes that share a similar expression pattern. We do this using the Boolean Query. The annotations for GUDMAP:9904 should be noted down.

From the annotations of Upk3a we can see it is expressed (present) in the collecting duct, medullary collecting duct and renal medulla and that it is not expressed (not detected) in the nephrogenic zone, renal cortex and maturing nephron. So, we’ll use this information to perform a Boolean Query to fins other genes with similar expression.

5. Using the Boolean Query
Navigate to

The user-interface the helps build a Boolean query is limited to only 3 components. Therefore, we’ll manually construct the query syntax we require. Help on using the Boolean Query syntax this is found at

Figure 5. Using the Boolean Query to find genes with similar expression

The syntax of the query we’ll use is:

GENE: p{in "collecting duct" TS17..TS28} AND p{in "medullary collecting duct" TS17..TS28} AND p{in "renal medulla" TS17..TS28} AND nd{in "nephrogenic zone" TS17..TS28} AND nd{in "renal cortex" TS17..TS28} AND nd{in "maturing nephron" TS17..TS28}

A total of 9 genes are returned by the query – including Upk3a (as expected).

6. Use Gene Expression Summaries to compare genes.

Each gene returned by the Boolean Query is displayed as a gene expression summary. This makes it possible to quickly view the expression of each gene and spot any similarities or differences.

Figure 6. Gene Expression Summaries. (A) Microarray Expression Profiles. (B) In situ Images. (C) Tick boxes to select genes for collections. (D) Collection features to build and modify user collections.

The Microarray Expression Profiles (Fig. 6A) can be compared for each gene to that of Upk3a (final row).
The in situ images for each gene can be viewed (Fig. 6B) by clicking on the ‘In situ expression images’ thumbnail.
Genes of interests can then be built into user collections by selecting genes (Fig. 6C) and using the collections features at the foot of the page (Fig. 6D). In order to be able to save collections you need to be logged in to the site. Logins can be obtained by e-mailing the GUDMAP Editorial Office (

By manually viewing expression profile, in situ images and in situ annotations of these genes we can further refine our list of genes.

Looking in more detail at the microarray expression profiles we can probably refine our list further. By clicking on the microarray expression profile for Upk3a we link to a more detailed microarray heatmap. If we select the Developing Kidney heatmap at the top of the page and click ‘Ok’ we can see where probes for Upk3a are expressed in components of the kidney. Upk3a shows expression in urothelium and medullary collecting duct from its microarray expression profile (Fig. 7).

Figure 7. Microarray Expression for Upk3a in the Developing Kidney

If we now look in more detail at the other genes in our list we can select those that share this type of expression. On this basis we can pick out AI836003, Foxa1, Gsdmc2 and Gsdmc4 as sharing microarray expression profiles similar to Upk3a based on elevated expression in both the urothelium and medullary collecting duct compared to other components of the developing kidney.

This set of genes represented potential disease candidate genes for Renal Adysplasia – generated through the use of the GUDMAP website to investigate gene expression patterns.

By using the GUDMAP Disease Resource we have identified a gene associated with a disease of interest. We have then been able to look at in situ expression entries in the GUDMAP database for that gene to see where it is expressed. Using this information we have performed a Boolean Query of the database to find other genes that share a similar expression pattern to our gene of interest.
From the results of this search we have then been able to further refine our results by looking in more detail at the microarray expression profiles of the gene.

This gives us a working set of potential disease candidate genes – simply by searching the gene expression data stored in GUDMAP.